English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Speaker-attributed automatic speech recognition (SA-ASR) is a task torecognize "who spoke what" from multi-talker recordings. An SA-ASR systemusually consists of multiple modules such as speech separation, speakerdiarization and ASR. On the other hand, considering the joint optimization, anend-to-end (E2E) SA-ASR model has recently been proposed with promising resultson simulation data. In this paper, we present our recent study on thecomparison of such modular and joint approaches towards SA-ASR on real monauralrecordings. We develop state-of-the-art SA-ASR systems for both modular andjoint approaches by leveraging large-scale training data, including 75 thousandhours of ASR training data and the VoxCeleb corpus for speaker representationlearning. We also propose a new pipeline that performs the E2E SA-ASR modelafter speaker clustering. Our evaluation on the AMI meeting corpus reveals thatafter fine-tuning with a small real data, the joint system performs 8.9--29.9%better in accuracy compared to the best modular system while the modular systemperforms better before such fine-tuning. We also conduct various error analysesto show the remaining issues for the monaural SA-ASR.

A Comparative Study of Modular and Joint Approaches for Speaker-Attributed ASR on Monaural Long-Form Audio